Talend executives vaunt their open source model

Talend CEO Bertrand Diard evangelizes his company’s open source approach to data management, depicting it as part of a revolution in enterprise software: data as a service to the fore.

Bertrand Diard is the CEO of Talend, an open source data management and application integration company founded in France; he is based in San Francisco. Yves de Montcheuil is vice president of marketing at the firm and is based in Lyon, France.

What follows is an edited version of interviews conducted with Diard and de Montcheuil by SearchDataManagement.co.uk, over the past month.

How do you see the market space you are in? What defines it? What are the dynamics?

Bertrand Diard: There is a fundamental change happening in the perspective of the [enterprise] software market. We see a huge differentiator on the model. Salesforce.com is an example of that. In terms of data, it will be considered more and more as a "as a service." We have moved on from considering [the mere] storage of data to finding the value in it. This is a huge change, from storage to finding the value proposition of data. The shift began three to five years ago. I’ve seen one estimate that the cost of storing one gigabyte of data in 2005 was $15.50; now it is 60 cents. If all you do is store, you are wasting your time. Now it is about getting the right level of quality and [determining the] management of it.


In terms of Talend’s value proposition, is it enterprise data management for organizations with constrained budgets?

Diard: Not really. We have Deutsche Post and Citibank as customers. It’s not about rationalizing your costs. It’s more related to the open source model than the technology. It is a new way to consume your software. The key thing is that we are based on standards.

In terms of competing on price, Oracle can provide, say, 80% discount [on their list price]. It’s not about the cost of the acquisition of the software; it’s about management into the future. Because we are open source, we can leverage Eclipse and Java, with lower development costs than our competitors, such as Informatica.


I’ve spoken to Talend customers who do find the migration path -- from free download to paying later, at their own pace, attractive. Is that your technology advantage?

Diard: Again, it is a matter of the model, not the technology. Look at the traditional model. You’ve got salespeople at IBM, for instance, who need to convince the CIO [chief information officer] to spend millions. In our case, the developers who download our products have their own skills. They start to prove the value proposition, [they talk to their bosses] and then they call us. It is the opposite approach. They are bridging the 20% [they can’t do] when calling us. It’s a disruptive model.

We provide choice -- that’s the most important thing. You can support yourself, fine; you are still providing value to our forum. And if at some point you need support from Talend, that’s fine too. We have had 15 million downloads since we were founded five years ago. We have a new download every 30 seconds. Usage is the most critical thing. And so it is a collaborative approach. When we release a new version, we have 5,000 people validating its quality. We could not afford to employ even 10% of those 5,000 developers.

Give me a sense of your customer base and revenue breakdown.

Diard: We have 150 average new paying customers per month and around 2,500 paying customers. In Europe, there is Orange, the French Ministry of Defence, BNP Paribas, Reed Exhibitions, SAB Miller, the Land Registry, Citibank, O2.There are others. Our business divides: 50% in Europe, 40% US, and 10% rest of the world; revenue is 70% from subscriptions, 30% from services. Of that 50% is training, 50% is architecture. We have around 400 employees in 12 offices.

But the main thing is there is no way we could have grown as we have done and had the recognition we have had from analysts on the traditional model. The model is viral. Indeed, our larger customers are often more knowledgeable than we are, which is a paradox. I am impressed every day with what the users do. The innovation driver is not that we are so brilliant; it’s the model.


How has the ongoing weak economic climate affected you?

Diard: It’s at times of economic difficulty, when companies and organizations change their habits. In the software market, the complexity of the global economy means you get changes for the future. Organizations reconsider all previous choices.

We have performed as we have because the software space is undergoing a big revolution. We had the Internet revolution between 1998 and 2002 and now software is leveraging the new capability of that. It is based on the cloud, but what is that? I really hate the hype around cloud. It is just your on-premises system abstracted out of your company. It is an evolution of the capability of the network, an evolution of infrastructure. You can deploy Talend on any server.

I’m not sure that everything will go to the cloud. We will see different combinations. The CIO challenge will be, say, managing SAP on-premises, Salesforce.com on a private cloud, and so on.


“Big data” is another topic du jour. How do you relate to that?

Diard: We are close to the Hadoop project. We embed directly on Hadoop, which is about processing a huge amount of data in parallel. eBay is a good example of one of our customers [using big data].


You recently announced an expansion of your Original Equipment Manufacturer (OEM) programme, which includes what you call the industry’s first embeddable 'big data technology,’ offered to software makers and Software as a Service providers. What does that mean?

Yves de Montcheuil: It means big data management that allows you not only to connect to big data sources -- such as machine-generated data, NoSQL databases and so on -- to get it into a structure such as Hadoop, but also to process it inside that structure. It is available to OEMs and direct customers. It saves a lot of time and energy in terms of doing the integration and quality on lots of data in a shorter time frame, not having to do it upstream.

Read more on Master data management (MDM) and integration